Dementia is a progressive syndrome of decline in cognitive function. For 5 million Americans with dementia, therapies slow progression but do not reverse or cure the disease. Nursing home care is common; 67% of people with dementia die in this setting. Shared decision-making about goals of care is the ethical standard for serious illness, yet families report poor quality communication, decision-making and palliative care. Compared with decisions about using or withholding a treatment, the goals of care approach encourages discussion and agreement on the primary goals of medical care, followed by treatment decisions designed to meet agreed upon goals. The study is a cluster randomized, controlled trial to test a decision support intervention for surrogates considering goals of care in advanced dementia. The intervention has two components: an audiovisual decision aid followed by a structured interdisciplinary care plan meeting. Study subjects are 300 nursing home residents with advanced dementia and their surrogates, recruited from 20 sites, to meet these Aims: Aim 1. To test the effect of the Goals of Care decision support intervention compared to an attention control on the quality of communication and decision-making, defined at 3 months follow-up as a) quality of communication; b) surrogate - health care provider concordance on goals of care; and c) surrogate report of treatment consistent with the resident's wishes. Aim 2. To test the effect of the intervention on quality of palliative care for residents with advanced dementia, defined at 6 months follow-up as a) number of palliative care domains addressed for the resident in the care plan; b) symptom management; and c) surrogate satisfaction with care for advanced dementia. Aim 3. To test the effect of the intervention on quality of dying for residents with advanced dementia, measured after death as a) surrogate - health care provider concordance on goals of care, and b) resident comfort in dying. This research will provide the first empiric test of the goals of care framework for dementia. It extends decision support research to surrogates, who make most decisions for patients with serious and incurable illness. To permit dissemination, the intervention design is pragmatic and well integrated with nursing home interdisciplinary care. PUBLIC HEALTH RELEVANCE: This cluster randomized trial compares an attention control to a Goals of Care decision support intervention on outcomes of quality of communication, quality of care and quality of dying. The trial will enroll 300 surrogate decision-makers for advanced dementia patients from 20 nursing homes.